Architecture operates as an evolving interface that organises energy flows and stabilises material forms. It shapes the built environment by merging form with energy, creating spaces that resonate with occupants through sensory experiences, and by establishing a thermodynamic relationship with the surroundings. This thesis examines how computational design, generative AI, and environmental awareness intersect in architectural creation. The focus lies on how intensive properties (such as light, temperature, and visual accessibility) interact with extensive attributes like physical dimensions to define space and structure. These ideas are integrated into broader discussions on computational design strategies, where generative and analytical methods work together, framing architecture as a practice that adapts to both environmental and human needs. Two case studies are at the core of this investigation. The first focuses on transitioning from intensive to extensive properties, using image-to-image translation to reverse-engineer design processes informed by environmental data. This method transforms environmental qualities into spatial forms, demonstrating how refined inputs can integrate these factors into architectural designs. The second case study reverses this approach, moving from extensive to intensive properties through text-to-image technologies and simulations. This process converts architectural forms into data, revealing attributes like embodied and operational energy, guiding architectural practices with AI-generated insights. The thesis introduces a framework that brings together generative and analytical processes, aligning creative and technical approaches to intertwine intensive and extensive properties within architectural design.